The PROC SQL Pass-Through Facility for Accessing and Querying External Databases

The PROC SQL Pass-Through Facility
DB2 Pass Through Query

FNAME
LNAME
CLAIMDT
CLAIMS
ANN
BECKER
01JAN1999:00:00:00
2003
CHRIS
DOBSON
13FEB2001:00:00:00
100
ALLEN
PARK
24SEP2001:00:00:00
10392
BETTY
JOHNSON
28NOV2001:00:00:00
3832
NANCY
PAUL
01JUN2001:00:00:00
1202
Presented by: Steve First

2997 Yarmouth Greenway Drive, Madison, WI 53711
Phone: (608) 278-9964 Web: www.sys-seminar.com
Systems Seminar Consultants, Inc

Topics:
Overview
Syntax
Options
Specifying Columns
Row Subsetting
Creating Views
Creating SAS Datasets
Date Constants and Conversion
Error Messages
PROC SQL Pass-Through Facility

The SQL Procedure Pass-Through Facility communicates with the DBMS
through the SAS/ACCESS engine.
The Pass-Through Facility allows you to:
pass native DBMS SQL statements to a DBMS

display the query results formatted as a report
create SAS datafiles and views from query results

The CONNECT clause allows you to communicate with the DBMS directly.
Syntax:
PROC SQL
<CONNECT TO dbms-name <AS alias><
<(connect-statement-argument-1=value
...<connect-statement-argument-n=value>)>>
<(dbms-argument-1=value
...<dbms-argument-n=value>)>>;
SELECT column-list
<=== SAS query
FROM CONNECTION TO dbms-name|alias
(select to from dbms-query)
<=== DBMS Query
optional PROC SQL clauses;
<DISCONNECT FROM dbms-name|alias;>
<QUIT;>
Notes:
The (DBMS-QUERY) is passed to the DBMS unchanged except for macros.

In a Pass-Through you specify:
which DBMS to connect to

DBMS passwords and options
a SELECT FROM CONNECTION TO:
a SELECT subquery that is passed directly to the DBMS enclosed in
parentheses
a DISCONNECT statement.
Notes:
The SELECT FROM CONNECTION allows you to specify SAS formats,
labels etc.
The subquery is passed to the DBMS unchanged except for macro
substitution.
PROC SQL Pass-Through Options For DB2 (partial)
SSID=DB2-subsystem-id
DB2 subsystem id.
SERVER=DRDA-server
DRDA server
AUTHID=authorization-id
DB2 authorization id
Example:
proc sql ;
connect to db2(ssid=ssc1);
. . .
Note:
These options are the same as the ones for the LIBNAME statement,
except they are coded in a different place.
A DB2 PROC SQL Pass-Through Example

Produce a report of all fields and 5 rows.
title 'DB2 Pass Through Query';
proc sql outobs=5;
connect to db2(ssid=ssc1);
select *
/*
from connection to db2
(select
*
/*
from ssctrain.benefits
/*
);
/*
%put &sysdbrc &sysdbmsg;
/*
disconnect from odbc;
quit;
sas select
*/
start pass-thru query*/

authid and table
*/
end pass-thru query */
return codes
*/
Notes:
Everything inside the parentheses refers to DB2 names, values. (BLUE)
Everything outside the parentheses is SAS. (RED)
Resulting Output
DB2 Pass Through Query
FNAME
LNAME
CLAIMDT
CLAIMS
ANN
BECKER
01JAN1999:00:00:00
2003
CHRIS
DOBSON
13FEB2001:00:00:00
100
ALLEN
PARK
24SEP2001:00:00:00
10392
BETTY
JOHNSON
28NOV2001:00:00:00
3832
NANCY
PAUL
01JUN2001:00:00:00
1202
Merging a DBMS Table with a SAS Dataset (DB2)

We read in a list of empcodes, most of which should match our employee
file.
libname db2lib db2 ssid=ssc1 authid=ssctrain;
data empcode;
input EMPNUMBER $3.;
datalines;
127
177
199
200
216
265
;
run;
<=== doesn't match
Merging a DBMS Table with a SAS Dataset (DB2)

Now merge empcode with the DB2 table.
data subemp;
merge db2lib.employee(in=onemp)
empcode(in=oncode);
by empnumber;
if oncode and onemp;
run;
proc print data=subemp;
title 'Selected Employee Records';
run;
Notes:
The entire DBMS table is returned to SAS (inefficient).
The DBMS will sort the EMPLOYEE table if necessary.
10
The Resulting Output
Selected Employee Records

Obs
1
2
3
4
5
FNAME
LNAME
JACK
CHRIS
LINDA
LYNN
BRENDA
KELLER
FOSTER
GILBERT
GOGGIN
WASHINGTON
STORENO
EMPNUMBER
33312
31381
33312
35618
31381
127
177
199
200
216
11
An Efficiency Problem
Merging a very small SAS file with a very large DBMS table can download
entire tables only to discard most rows because they don't match.
Is there some way to tell SAS, only to bring down matching rows?
12
The DBKEY Option

Read a second file only if keys match.
Syntax:
data sasdataset;
set sasdataset1;
set sasdataset2(dbkey=index) key=dbkey;
Notes:
Read every row of sasdataset1, then directly read sasdataset2 by a key

value.
sasdataset1 must contain key values to locate in sasdataset2.
_IORC_ is set to 0 if matching row is found in sasdataset2.
13
The DBKEY Option

Match the two files from the previous example.
options errors=0;
/* don't log not found records */
data subemp;
set empcode;
set db2lib.employee(dbkey=empnumber)
key=dbkey;
if _iorc_=0 then
/* match ? */
output;
/* output */
run;
proc print data=subemp;
title 'Selected Employee Records';
var fname lname storeno empnumber;
run;
Notes:
Only matching records are retrieved from the DBMS table (efficient).
The DBKEY variable may or may not be an index on the DBMS.
14

Obs
1
2
3
4
5
FNAME
LNAME
JACK
CHRIS
LINDA
LYNN
BRENDA
KELLER
FOSTER
GILBERT
GOGGIN
WASHINGTON
STORENO
EMPNUMBER
33312
31381
33312
35618
31381
127
177
199
200
216
Note:
DBKEY should only be used when SAS dataset is very small.
15
Joining a DBMS Table with a SAS Dataset (DB2)

Can we join the EMPCODE SAS dataset to EMPLOYEE table with PROC
SQL?
Yes, gather the EMPCODE dataset as before.
libname db2lib db2 ssid=ssc1 authid=ssctrain;
data empcode;
input EMPNUMBER $3.;
datalines;
127
177
199
200
216
265
;
run;
<=== doesn't match
16
Joining a DBMS Table with a SAS Dataset (DB2)

(continued)
Join the EMPLOYEE table with the EMPCODE SAS dataset.
proc sql;
create table work.subemp as
select a.fname,
a.lname,
a.storeno,
a.empnumber
from db2lib.employee a,
work.empcode b
where a.empnumber = b.empnumber
;
quit;
proc print data=work.subemp;
. . .
Notes:
The entire DBMS table is sometimes downloaded to SAS (expensive)
17

Obs
1
2
3
4
5
FNAME
LNAME
JACK
CHRIS
LINDA
LYNN
BRENDA
KELLER
FOSTER
GILBERT
GOGGIN
WASHINGTON
STORENO
33312
31381
33312
35618
31381
EMPNUMBER
127
177
199
200
216
18
DBKEY Passes Keys of Rows to Select

DBKEY can also be used with PROC SQL to download only matching rows.
proc sql;
create table work.subemp as
select a.fname,
a.lname,
a.storeno,
a.empnumber
from db2lib.employee(dbkey=empnumber) a,
work.empcode b
where a.empnumber = b.empnumber
;
quit;
proc print data=work.subemp;
. . .
Notes:
This is much more efficient when SAS dataset is small.
19

Obs
1
2
3
4
5
FNAME
LNAME
JACK
CHRIS
LINDA
LYNN
BRENDA
KELLER
FOSTER
GILBERT
GOGGIN
WASHINGTON
STORENO
33312
31381
33312
35618
31381
EMPNUMBER
127
177
199
200
216
20
Debugging and Optimizing Queries

OPTIONS SASTRACE and SASTRACELOG will show generated SQL.
Examples:
options sastrace=',,,d';
options sastraceloc=saslog;
or
options debug=dbms_select;
Notes:
debug=dbms_select shows only the select statements generated.
21
Debugging and Optimizing Queries (continued)

Sort a MSA table and keep only the first and last name.
libname msalib odbc dsn=mdbtest;
options sastrace=',,,d' sastraceloc=saslog;
proc sort data=msalib.benefits
out=benefit2(keep=lname fname) nodupkey;
by lname fname;
run;
proc print data=benefit2;
title 'Benefit2';
run;
22
The Generated Log

The trace doesnt contain much of interest on this page.
994
995
996
997
proc sort data=msalib.benefits

out=benefit2(keep=lname fname) nodupkey;
by lname fname;
run;
TRACE: Successful connection made, connection id 4 4

1342061696 no_name 0 SORT
TRACE: Database/data source: mdbtest 5 1342061696
no_name 0 SORT
TRACE: AUTOCOMMIT is NO for connection 4 6 1342061696
no_name 0 SORT
TRACE: Using FETCH for file benefits on connection 4 7
TRACE: Change AUTOCOMMIT to YES for connection id 4 8
1342061696 no_name 0
23
The Generated Log (continued)

All of the columns were selected.
TRACE: SQL stmt execute on connection 4:
SELECT * FROM
`benefits` 9
TRACE: SQL stmt prepared on statement 0, connection 0
is: SELECT `LNAME`, `FNAME`, `CLAIMDT`, `CLAIMS`
FROM `benefits` 10 1342061696 no_name 0 SORT
TRACE: DESCRIBE on statement 0, connection 0. 11
NOTE: 13 observations with duplicate key values were

deleted.
NOTE: There were 39 observations read from the data set
MSALIB.benefits.
How could this be written more efficiently?
24
A Better Program
Keeping only the columns needed as an input option will result in a shorter
query.
libname msalib odbc dsn=mdbtest;
options sastrace=',,,d' sastraceloc=saslog;
proc sort data=msalib.benefits(keep=lname fname)
out=benefit2 nodupkey;
by lname fname;
run;
proc print data=benefit2;

title 'Benefit2';
run;
25

The PROC SQL Pass-Through Facility for Accessing and Querying External Databases

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The PROC SQL Pass-Through Facility for Accessing and Querying External Databases

Uploaded by

Copyright:

Available Formats

The PROC SQL Pass-Through Facility

DB2 Pass Through Query

Presented by: Steve First